Proc. Natl. Acad. Sci. USAVol. 92, pp. 10374-10378, October 1995Microbiology

Genetic rearrangements in the rjb regions of Vibrio cholerae 01and 0139

(chromosomal rearrangements/serotype/lipopolysaccharide/H-repeat/insertion sequences)

UWE H. STROEHER*, KATHY E. JEDANI*, B. KATE DREDGE*, RENATO MORONA*, MELISSA H. BROWN*t,LITSA E. KARAGEORGOS*t, M. JOHN ALBERT§, AND PAUL A. MANNING*¶*Microbial Pathogenesis Unit, Department of Microbiology and Immunology, University of Adelaide, South Australia 5005, Australia; and §International Centrefor Diarrhoeal Diseases Research, Bangladesh, Dhaka, Bangladesh

Communicated by Allen Kerr, Adelaide, Australia, May 16, 1995

ABSTRACT The recent emergence of a pathogenic newnon-01 serotype (0139) of Vibrio cholerae has led to numerousstudies in an attempt to identify the origins of this new strain.Our studies indicate .that 0139 strains have clear differencesin the surface polysaccharides when compared with 01strains: the lipopolysaccharide can be described as semi-rough. Southern hybridization with the 01 rjb region dem-onstrates that 0139 strains no longer contain any of the rjbgenes required for the synthesis of the 01 0-antigen or itsmodification and also lack at least 6 kb of additional contig-uous DNA. However, 0139 strains have retained rfaD andhave a single open reading frame closely related to three smallopen reading frames of the 01 rjb region. This region is closelyrelated to the H-repeat of Escherichia coli and to the trans-posases of a number of insertion sequence elements and hasall the features ofan insertion sequence element that has beendesignated VcISI. Transposon insertion mutants defective in0139 0-antigen (and capsule) biosynthesis map to the samefragment as VcISI. Preliminary sequence data of comple-menting clones indicate that this DNA encodes a galactosyl-transferase and other enzymes for the utilization of galactosein polysaccharide biosynthesis. We propose a mechanism bywhich both the Ogawa serotype of 01 strains and the 0139serotype strains may have evolved.

Cholera in humans has traditionally been associated withVibrio cholerae serotype 01. However, in 1992, apparently forthe first time in recorded history, a non-O1 strain of V choleraewas responsible for an epidemic of cholera (1, 2). Previously,non-O1 strains caused only sporadic cases of diarrhea but notepidemics of cholera (3). This strain did not react withantibodies to any of the known serotypes and was designated0139 synonym Bengal (4) for its first site of isolation. By June1993, this serotype had reached Karachi and was displacing 01as the epidemic serotype. A similar 0139 strain may haveemerged in Argentina (5).Over 100,000 cases of 0139 cholera have been reported, and

the fear exists that 0139 could initiate the next, eighthpandemic. How has this new serotype arisen? It has beenproposed that the 0139 (Bengal) serotype is due to a mutationin the 0-antigen (rjb) locus (6). Our data do not support sucha simple hypothesis.

Previous studies have sought to distinguish 0139 from 01strains by means other than serotyping. Basically these studiessuggested that 0139 (Bengal) is very closely related to the ElTor biotype 01 strains (7-10). However, unlike 01 strains,0139 strains are encapsulated (8, 11).

Using oligodeoxynucleotides as probes for the 01 rfbR andrfbS genes, Johnson et al. (11) suggested that these genes areabsent in 0139 strains. However, in this study we find that both

01 and 0139 strains have a region defined by the rJbQ, -R, and-S genes (12) that is related to an open reading frame (ORF)called the H-repeat of RHS elements in Escherichia coli. RHSelements were first recognized as chromosomal rearrangementhotspots (13, 14). This region appears to represent an insertionsequence (IS).II

MATERIALS AND METHODSBacterial Strains. The V cholerae 01 strains used included

017 (El Tor, Ogawa); 569B (classical, Inaba); AA14073 (ElTor, Ogawa); BM69 (El Tor, Inaba); 64 (El Tor, Ogawa); andH-1 (El Tor, Ogawa.) Strains 1074 and 1196 are nontoxigenicEl Tor 01 environmental isolates. AI-1837, AI-1838, AI-4450,AI-1841, AI-1852, AI-1854, AI-1855, and AI-4260 are 0139isolates from the Indian subcontinent. Arg-3 is an 0139 isolatefrom Argentina (5). The rough strains V663 (rjb::Tn5) andV665 (rfl::TnS) are derived from 569B (15).

Southern Hybridization. DNA from V cholerae 01 and0139 strains was prepared as described (16). Transfer ofDNAfrom agarose gels to nitrocellulose filters (Schleicher &Schuell) was performed as described by Southern (17) withmodifications (18). Detection was by enhanced chemilumines-cence (ECL; Amersham) and autoradiography.

Polymerase Chain Reaction (PCR). Amplification usingAmplitaq DNA polymerase (Hoffmann-La Roche) was car-ried out by standard protocols with the oligodeoxynucleotideprimers described.

Antibodies. The initial polyclonal antiserum (antiserum 1)and monoclonal antibody (mAb) 1CL-12 have been described(19). Antiserum 2 to 0139 was prepared from a rabbitimmunized with heat-killed AI-1837 and subsequently wasabsorbed to remove crossreactive antibodies.

Lipopolysaccharide (LPS) Silver Staining and Immuno-blotting. LPS from 1-ml overnight (18-hr) cultures (20) wasanalyzed by SDS/PAGE followed by silver staining or immu-noblotting (21).

Sequencing Procedures. The nucleotide sequence of the rjbregions of 017 and 569B was determined as described (12).Sequencing of the 0139 region was performed with an AppliedBiosystems 373A automated DNA sequencer. Sequences canbe obtained under the accession numbers X59554 (Ogawa)and X91246 (0139).T7 Expression System. The temperature-inducible, T7 RNA

polymerase/promoter system (22) was used to express

Abbreviations: IS, insertion sequence; LPS, lipopolysaccharide; mAb,monoclonal antibody; ORF, open reading frame.tCurrent address: School of Biological Sciences, University of Sydney,Sydney, NSW 2006, Australia.*Current address: Division of Clinical Virology, Institute of Medicaland Veterinary Science, Adelaide 5001, Australia.1To whom reprint requests should be addressed.11 The sequences reported in this paper have been deposited in theGenBank database (accession nos. X91246).

10374

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

May

25,

202

0

Proc. Natl. Acad. Sci. USA 92 (1995) 10375

L-[35S]methionine (Amersham)-labeled proteins. Proteinswere visualized by autoradiography following SDS/PAGE.Molecular size markers were from Pharmacia.TnphoA Mutagenesis. The suicide plasmid pRT733 (23) was

used for isolation of TnphoA insertions in strain AI-1837 byselecting kanamycin- and polymyxin B-resistant exconjugants.

RESULTS

SDS/PAGE of the LPS of 0139 Strains. We examined anumber of V cholerae 0139 (Bengal) strains used in otherstudies (8, 24). Analysis of LPS from these strains by SDS/PAGE and silver staining showed that 0139 strains do notcontain the typical ladder of LPS bands detected in 01 strains,nor do they resemble typical rough 01 mutants. However,0139 strains do have a distinct pattern resembling that of asemirough LPS: a region corresponding to the lipid A-coreoligosaccharide of 01 is observed, as well as a more heavilystaining, slower-migrating region (Fig. 1 Left). From theapparent molecular size, the difference between the two 0139bands is likely to be only one or two sugars which presumablycap the lipid A-core. The LPS from the 0139 Argentinianisolate (Arg-3) differs again and is slightly smaller.

Colony blotting using 0139 mAb 1CL-12 (19) revealedstrong crossreactivity with rough 01 strains (V663 and V665),indicating that the mAb reacts with epitopes present in thecores of both 01 and 0139 LPS (data not shown). Westernblotting of proteinase K-treated whole cells with antiserum 2showed crossreactivity to the slower-migrating material of bothBengal and Argentinian 0139 strains (Fig. 1 Right).

Analysis of the rjb Region of 0139 by Southern Hybridiza-tion. The unusual LPS of 0139 strains led us to initiallyinvestigate the rib region in the 0139 serotype by probing withfour plasmids which span the 01 rjb region (Fig. 2). ProbespPM4202 and pPM4203 did not hybridize with genomic DNAfrom the 0139 strains, but pPM4201 and pPM4204 hybridizedweakly. Probes derived from rfaD hybridized strongly to all0139 strains, including Arg-3 (data not shown). The use of therfaD probe on EcoRI-digested chromosomal DNA revealed arestriction polymorphism within the rfaD genes of 0139 strains

&0

- l~

n

CU2CQNtR!

LC

cu ca..o

u)

t-

't CU n2 aD cr C

bkOI CD 2s! !Q E -- COc w co n

1

jk I I D tI-

0

Mod.A-core

A-core

C m. 4.

.h

FIG. 1. (Left) Analysis of LPS from V cholerae 01 and 0139 bySDS/15% PAGE with silver staining. Strains are indicated above thelanes. V663 is an rfb::TnS mutant of V. cholerae 01 (569B) and lacksany 0 antigen (15). V942[pPM4225] is the complemented V942 strain.The 0139 transposon mutant is V942. The positions of the lipid A-coreand 0-antigen linked to lipid A-core are indicated, as is the positionof the modified lipid A-core in 0139 strains. (Right) Western blot ofan identical gel, probed with the 0139-specific antiserum 2. Theantibodies react exclusively with the substituted lipid A core.

distinguishing the Bengal isolates from the Argentinian strain.In addition, PCR analysis revealed that Arg-3 has a defectiverfaD gene. These data correlate with the altered LPS patternseen in Arg-3 (Fig. 1). By PCR using primers within rfaD andrbA, we showed that one end of the deletion in the 0139Bengal strains is close to rfaD but that the deletion endpointin Arg-3 is within rfaD (data not shown).

Southern hybridization analysis with the probe derived fromthe PCR product generated with primer 773 (corresponding tothe 17-bp repeat sequences flanking the rJbQRS DNA; Fig. 3)revealed only a single copy of this region in 01 strains.However, two distinct copies on separate Sac I fragments of21.5 and 6.8 kb were detected in the 0139 strains. Only one ofthese copies contains a distinctive EcoRI restriction polymor-phism and can be PCR amplified, indicating that the secondcopy is incomplete or is only a related, region of DNA.

Plasmid pEVX12 (25), which contains DNA outside of theSac I fragment harboring the 01 rjb region, was used to probe0139 chromosomal DNA. The probe hybridized only weakly,probably with the rJbQRS region, suggesting that more thanjust the 01 rJbA-P and rJbT genes are absent in 0139 strains.pEVX12 contains two ORFs (ORFI and ORFII) not previ-ously described which show homology to 0-antigen biosyn-thesis proteins (P.A.M., A. Fallerino, and C. Mavrangelos,unpublished work).Given the complete nucleotide sequences of the El Tor

Ogawa and classical Inaba 01 rjb regions, we assessed theextent of homologous DNA in 0139 strains. Only the rfaD andribQRS regions were detected. Together with the data from thesubclones and pEVX12 (Fig. 2), this indicates that >20 kb of01 DNA is not present in 0139 (Bengal).

rJbQ, -R, and -S Are Homologous to E. coli H-Repeats. Thepredicted amino acid sequences of the 01 rflQ, -R, and -SORFs share similarity to the single ORF found within theH-repeat of RHS elements of E. coli (14). These H-repeats arealso bounded by inverted repeats as is the rJbQRS region.Immediately adjacent to the 3' end of the repeat is thepromoter for the Ogawa determining gene, rflT (12). As willbe seen below, rfbQ, -R, and -S form a single ORF in 0139 thatis homologous not only to the H-repeat of E. coli but also tothe transposases of a number of IS elements. Thus, the entireregion found in 0139 is organized like an IS element, which wehave designated IS1358; the defective version found in 01strains is IS1358dl.Sequence of IS1358. The PCR product obtained from 0139

strain AI-1837 by using primer 773 has been sequenced (Fig.3) and found to comprise a single ORF and not three separateORFs as are found in the 017 01 strain. Thus, IS1358 of 0139resembles the H-repeat of the E. coli RHS B and E elements,which are the H-repeats in E. coli containing a single ORF(14). The IS1358 ORF also most closely resembles the ISelement (AsIS1) found inAeromonas salmonicida (26). This isof particular interest since AsIS1 is the first of these elementswhich has been shown to actually transpose. The transposasesof E. coli IS elements show a lower level of identity (- 18-20%)than AsIS1 (27.5%). Thus, it appears likely that the ORFwithin IS1358 in 0139 encodes a transposase and that togetherwith the inverted repeats the whole region represents an ISelement. -

Restriction Fragment Polymorphisms Within IS1358. TheribQRS region (IS1358 element) was PCR amplified from allavailable 0139 strains and from several 01 strains. Restrictionenzyme analysis revealed an EcoRI polymorphism: all 0139strains have an EcoRI site "400 bp from one end of theinverted repeat, but this is not present in any of the 01 strainstested (Figs. 3 and 4). This region of the chromosome has beena site of genetic rearrangement activity during the evolution of0139.The New LPS Genes in 0139 Are Linked to IS1358. Trans-

poson mutagenesis of the 0139 strain AI-1837 using TnphoA

Microbiology: Stroeher et al.

Dow

nloa

ded

by g

uest

on

May

25,

202

0

10376 Microbiology: Stroeher et al. 1

O- l-- *- *- *. *0 -4n &OA Al AlsA c A IA A A lAI

-I Ima~~ II

0r^D

5 10 15A B D E G HI K L U N OPQRS TORFB

Proc. Natl. Acad. Sci. USA 92 (1995)

m m _ 0120ORFA

pPM4201

pPM4202pPM4203

pPM4204pEVX12

IfxD EIiiZA L T || ORFII || ORFI I

+ 179/277 -_ -

+ 208/184+ 171/184+ 167/184+ 205/292+ 910/185+ 171/1058

910/1057

FIG. 2. The rjb region of V cholerae 01. The rib genes are indicated by boxes rfbA-T. ORFI and ORFII encode proteins homologous to otherpolysaccharide biosynthetic genes. Plasmids pPM4201, pPM4202, pPM4203, and pPM4204 were used as probes. The PCR probe 773 was obtainedby using a single oligodeoxynucleotide primer, no. 773, to the inverted repeat flanking VcISl. PCR probe 434/270 was derived fromoligodeoxynucleotide primers 434 (outside of rfaD) and 270 (within rfaD). PCR amplification was carried out with the indicated primer pairs. Positivehybridization with 0139 genomic DNA by either Southern hybridization or PCR amplification is indicated by +. Negative reactions are indicatedby -.

(23) yielded two mutants, V942 and V946, which no longerreacted with antiserum 2. Mutant V942 was selected forfurther analysis by silver staining and shown to lack theslower-migrating 0139 LPS band and to be defective in capsulebiosynthesis (data not shown) (Fig. 1). Southern hybridizationusing the PCR product from primer 773 as a probe for IS1358revealed that the transposon had inserted within the same SacI fragment (Figs. 4 and 5), indicating that the new LPS genesare linked to IS1358. V942 can be complemented to wild typeby a cosmid clone (pPM4225) derived from the 0139 strainAI-1837 (Fig. 5). Sequencing of this cosmid has shown that itcontains genes homologous to known LPS genes associatedwith incorporation of galactose, including a galactosyltrans-ferase gene (U.H.S. and P.A.M., unpublished work).

Expression of the IS1358 ORF. The IS1358 ORF is pre-dicted to encode a 42-kDa protein. We have used a T7 RNApolymerase/promoter system (22) to express this region ofDNA from both 01 and 0139 strains. A protein whichmigrates at -42 kDa (Fig. 6) is encoded within the 0139region, but no product is seen with 01. Thus, IS1358 appearsto contain an intact transposase gene, but this is defective in01 strains.

DISCUSSION

Our SDS/PAGE analysis shows that strains of the 0139serotype have LPS which differs from LPS from 01 strains.The LPS of the 0139 strains does not appear to have typical

long 0-antigen chains, unlike 01 strains, where the core issubstituted with an average of 17 repeat units. The 0139 LPShas the appearance of an efficiently substituted core oligosac-charide, albeit with only one or two additional sugar molecules.This is reminiscent of the semirough LPS or lipooligosaccha-ride of organisms such as Neisseria. We propose that the 0139lipid A-core is either the same or closely related to that of 01strains, since a mAb capable of recognizing the 0139 LPS (19)showed marked crossreactivity with rough 01 strains, but notwith smooth 01 strains.Our Southern hybridization and PCR analyses show that the

change from 01 to 0139 is not just a simple mutation in therjb region but a complete substitution of this region. Further-more, Southern hybridization shows that the deletion extendsbeyond the rfb region.The events involved in the change from 01 to 0139 are not

fully elucidated. One possible scenario for the evolution of0139 is that IS1358, which we have shown to be linked to thenovel 0-antigen genes, was able to facilitate their incorpora-tion by recombination into the V cholerae 01 chromosome viahomology to rfbQRS (IS1358dl). IS1358 and IS1358dl are

96% homologous, sufficient to permit homologous recombi-nation, and identical in size. Following the acquisition of thenovel LPS genes, one might expect incompatibility with theoriginal 01 rfb genes that, with immune pressure, would selectfor the 0139 genes. Similarly, the loss of ORFI and ORFII mayalso have been due to incompatibility with the incoming 0139capsular polysaccharide genes and thus have led to their

+1

+1-/1

PCR 773 probe +PCR 434/270 probe +

_ -.0_ _

_.0 14

_.0 _-

_40 _-

434/271-502/271 -

434/270-*502/270-.102/270197/270211/2702U/236102/238197/238237/238237/210237/295197/909

_. _-

_. _-

-0-

-10 4--

-0.

-.& 41--

-.41. 4--

Dow

nloa

ded

by g

uest

on

May

25,

202

0

Microbiology: Stroeher et al.

TACCAATGCCTTCCTTAATCTCAGTTTI'rGATGTAACCCGATG;AAATrrAAGATCACCTTRBS

GTAGATCATTTCAGTTGTTTGTAATAGCTCCTAGACAATATAGGAGCCTAAATATGAGC--- ---- M S

Proc. Natl. Acad. Sci. USA 92 (1995) 10377

59

119

1792

GAGTTAATCAACCCATTTATGCATTTCCAAATCATCAAAGACTATCGACAAGAAAGCAAA 239E L I N P F M H F Q I I K D Y R Q E S K 22

GTAGAACACAAATTATCAGACATTATTT¶IlACAA¶IT'IGTGGTGTIrGTCGTGTCAC 299V E H K L S D I I L L T I C G V L S C H 42

GATGGC GGGATGGCATAGATTTCCTTAAACGATAT 359D G W D G I I D F G N A R L D F L K R Y 62

EcoRIC_ 419

G H F E A G I P S A D T L S R V M G M IHindIII

AATCCIGTCGCTTTGCAAAGAGCTTCATTGCCTGGATGAGAACTGCCATACACTGACGN P V A L Q R S F I A W M K N C H T L T

r--m

CV1

_.L.

82

479102

GAiGTGAAGTGATTccATcGAcGGTAAAACATTAcGCGGCTcTTACGACCGcTCGAAA 539D G E V I A I D G K T L R G S Y D R S K 122

GGCAAAGGAACGATCCATATGGTGAACGCVITTTCTACAC GAATGAGTGGG 599G K G T I H M V N A F A T A N G M S I G 142

CAACTGAAGGTTGATTCAAGAGTAACGAGATTACCGCGATCCCCAAGCTACTTGACTTG 659Q L K V D S K S N E I T A I P K L L D L 162

CTAGATGTAAAAGGccTTGATTAcGATrGATccATGGGCTGTCAAAAGAAAATAGcG 719L D V K G C L I T I D A M G C Q K K I A 182

CAGAAAATCCGTGATAAAGAAGCGGATTATTTATTGGOGGTCAAAGGCAATCAGGGAATG 779Q K I R D K E A D Y L L A V K G N Q G M 202

CTTGAGCAAGCCTTTGATGATTATTTTCGAATGGACATGCTTCAAGACTTCGACGGTAGT 839L E Q A F D D Y F R M D M L Q D F D G S 222

TCA AG G A T A 899S Y S T Q E K S H G R I E T R V A L V N 242

CGCGAGGT?GTCGGTATAP TGAACATGAATGGCCTCTGGGC 959R D L S V L G D I E H E W P G L K S M G 262

ATCGTGGCTTCAATTCGACAAGAATCGGCAGTAGCAACAGAGCAAGATGTGAGTATTIGT 1019I V A S I R Q E S A V A T E Q D V S I R 282

TACTACATACGITCTCAC 1079Y Y I C S K E L E A Q T L L E A T R S H 302

TATCATTGGTC 1139W G V E V M H W S L D T A F C E D N S R 322

ATTAGAGCGGACGATCGAGCAGAGGCATGCAAGGATCAGGCAGATA1GTTrGAACCTA 1199I R A D D R A E A F A R I R Q I C L N L 342

TTAAAGAGCGAAACAGA T T 1259L K S E T T F K G G I K R K R M N C A M 362

GACGAAAAGTAC G AC CT ATGCGGT 1319D E K Y L S K V L E S L T * 376

TTCGTG 1326

FIG. 3. Nucleotide sequence of IS1358 from 0139 strain AI-1837.The sequence corresponds to IS1358 within the 17-bp inverted repeatsat the ends (solid arrows). Smaller inverted repeats (dashed arrows)cover the ribosome binding site (RBS), which is shown in boldfacetype. The stop codon is indicated by an asterisk. Relevant restrictionsites are shown and underlined. The EcoRI site corresponding to apolymorphism between 01 and 0139 strains is indicated.

elimination. If IS1358 mediated the event, the novel DNAcould be expected to be flanked by two copies of IS1358, andindeed there appear to be two elements located on adjacentSac I fragments of 21.5 kb and 6.8 kb in 0139 strains (data notshown). Since the second copy can no longer be amplified byPCR and does not contain the EcoRI site, it may well be theremains of the 01 rfbQ, -R, and -S genes after the recombi-nation and deletion events.

FIG. 5. Southem hybridization of Sac I-digested 01 (H-1) and0139 (AI-1837) genomic DNA with probe 773. Plasmid pRMB2contains the 20-kb fragment harboring the 01 rjb region and was usedas a positive control for IS1358dl. DNA fragments were electropho-resed in 0.8% agarose, transferred to nitrocellulose, and probed withPCR probe 773, which spans IS1358dl. The IS1358dl genes lie on a20-kb Sac I fragment in the 01 strain (H-1), whereas in the 0139isolate (AI-1837) homologous DNA lies on a 21.5-kb Sac I fragment.This fragment has an increased size in the TnphoA mutant of AI-1837(V942), which lacks 0 antigen. This increase corresponds to the sizeof the transposon insert.

An alternative explanation for the origin of the 0139 regionis that the segment of 0139-specific DNA may have beenrecombined via flanking homologous DNA such as rfaD at theproximal end and whatever the corresponding region is at thedistal end. This could be a simple transduction event, not asmechanistically complex as that above.

Since IS1358 shows strong homology to Aeromonas AsIS1,a known mobile genetic element (26), it is possible that IS1358is also an IS element. Further evidence, relating to organiza-tional features, suggests that IS1358 is an IS element: IS1358is flanked by inverted repeats (16 of 17 nt are identical) at itsends and has an inverted repeat encompassing the ribosomebinding site. A similar genetic structure is seen in IS10 andIS50, where these inverted repeats have been shown to stopreadthrough transcription from external promoters and thusreduce the expression of the transposase (27, 28). If IS1358 is

o L: CLo

cq co cQ C\ E-I

1 -0: )

kDa4 94

4 67

4 43

4 30

la

4 20.1

4 14.411358 F O 1 0139 Core Modification Genes

TnphoA lkbV942

FIG. 4. Model for the organization of the chromosomal region inthe vicinity of IS1358 in 0139. The 0139 physical map of the 21-kb SacI fragments containing IS1358 is based upon Southern hybridizationdata obtained from the strain AI-1837. Locations of the new coremodification genes were determined by transposon insertion analysiswith TnphoA and nucleotide sequencing.

FIG. 6. Expression of the IS1358-encoded protein. Whole cellsharboring the indicated plasmids were induced to overexpress theprotein and solubilized in sample buffer. After SDS/15% PAGE,proteins were stained with Coomassie blue G250. pPM4250 andpPM4251 contain IS1358dl cloned in opposite orientations intopGEM-T (Promega). pPM4252 contains the IS1358 element clonedinto pGEM-T under the control of the T7 promoter; pPM4253 has thesame clone in the opposite orientation.

am

Dow

nloa

ded

by g

uest

on

May

25,

202

0

10378 Microbiology: Stroeher et al.

confirmed to contain an IS, it will represent the first IS elementdetected in V. cholerae and only the second mobile geneticelement, after RS1 (29).From a map of the 0139 chromosome around IS1358 based

on Southern data (Fig. 4), it appears that the change from 01to 0139 has involved not only the replacement of the originalrJb region with a new one but also the removal and replacementof -6 kb of additional contiguous DNA.Given that IS1358dJ of 01 represents a defective form of

IS1358, the Ogawa form of 01 strains may have derived froman Inaba ancestor by IS1358-mediated introduction of rJbT(the Ogawa determining gene) and degenerated due to alack of selection. In support of this notion, variation in theIS1358dl ORF can be detected in different 01 strains (ref.12; unpublished data). The role of H-repeat-related ele-ments in generating new 0-antigen or LPS specificities hasbeen proposed by Xiang et al. (30) and may represent acommon mechanism. Further studies of these elements mayprovide insights into horizontal gene transfer and geneticevolution of 0-antigen specificities in Gram-negative bac-teria.The polymorphisms located in the vicinity of rfaD can be

used to readily distinguish the Argentinian strain from theBengal 0139 strains. Indeed, there are still further differencesin Arg-3 to both 01 and 0139 Bengal strains which suggest thatthis isolate has arisen independently and from a differentprogenitor.

We acknowledge the support of the Australian Research Counciland the Diarrheal Diseases Global Vaccine Program of the WorldHealth Organization.

1. Albert, M. J., Siddique, A. K., Islam, M. S., Faruque, A. S. G.,Ansaruzzaman, M., Faruque, S. M. & Sack, R. B. (1993) Lancet341, 704.

2. Ramamurthy, T., Garg, S., Sharma, R., Bhattacharya, S. K., Nair,G. B., Shimada, T., Takeda, T., Karasawa, T., Kurazono, H., Pal,A. & Takeda, Y. (1993) Lancet 341, 703-704.

3. Morris, J. G., Jr. (1990) Epidemiol. Rev. 12, 179-191.4. Shimada, T., Nair, G. B., Deb, B. C., Albert, M. J., Sack, R. B. &

Takeda, Y. (1993) Lancet 341, 1347.5. Rivas, M., Toma, C., Miliwebsky, E., Caffer, M. I., Galas, M.,

Varela, P., Tous, M., Bru, A. M. & Binsztein, N. (1993) Lancet342, 926-927.

6. Hall, R. H., Khambaty, F. M., Kothary, M. & Keasler, S. P.(1993) Lancet 342, 430.

7. Das, B., Gosh, R. K., Sharma, C., Vasin, N. & Gosh, A. (1993)Lancet 342, 1173-1174.

8. Johnson, J. A., Albert, M. J., Panigrahi, P., Wright, A. C., Joseph,A., Comstock, L., Truksis, M., Michalski, J., Johnson, R. J.,

Kaper, J. B. & Morris, J. G., Jr. (1993) in Proceedings of the 26thJoint Conference, U.S.-Japan Co-operative Medical Science Pro-gram, Cholera and Related Diarrheal Diseases (National Institutesof Health, Bethesda, MD), pp. 35-38.

9. Calia, K. E., Murtagh, M., Ferraro, M. J. & Calderwood, S. B.(1994) Infect. Immun. 62, 1504-1506.

10. Waldor, M. K. & Mekalanos, J. J. (1994) Infect. Immun. 62,72-78.

11. Johnson, J. A., Salles, C. A., Panigrahi, P., Albert, M. J., Wright,A. C., Johnson, R. J. & Morris, J. G., Jr. (1994) Infect. Immun. 62,2108-2110.

12. Stroeher, U. H., Karageorgos, L. E., Morona, R. & Manning,P. A. (1992) Proc. Natl. Acad. Sci. USA 89, 2566-2570.

13. Lin, R.-J., Capage, M. & Hill, C. W. (1984)J. Mol. Biol. 177, 1-18.14. Zhao, S., Sandt, C. H., Feulner, G., Vlazny, D. A., Gray, J. A. &

Hill, C. W. (1993) J. Bacteriol. 175, 2799-2808.15. Ward, H. M. & Manning, P. A. (1988) Mol. Gen. Genet. 218,

367-370.16. Manning, P. A., Heuzenroeder, M. W., Yeadon, J., Leavesley,

D. I., Reeves, P. R. & Rowley, D. (1986) Infect. Immun. 53,272-277.

17. Southern, E. M. (1975) J. Mol. Biol. 98, 503-517.18. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular

Cloning: A Laboratory Manual (Cold Spring Harbor Lab. Press,Plainview, NY), Vols. 1-3, 2nd Ed.

19. Qadri, F., Chowdhury, A., Hossain, J., Chowdhury, K., Azim, T.,Shimada, T., Islam, K. M. N., Sack, R. B. & Albert, M. J. (1994)J. Clin. Microbiol. 32, 1589-1590.

20. Hitchcock, P. J. & Brown, T. M. (1983)J. Bacteriol. 154,269-277.21. Macpherson, D. F., Morona, R., Beger, D. W., Cheah, K.-C. &

Manning, P. A. (1991) Mol. Microbiol. 5, 1491-1499.22. Tabor, S. & Richardson, C. C. (1985) Proc. Natl. Acad. Sci. USA

82, 1074-1078.23. Taylor, R. K., Manoil, C. & Mekalanos, J. J. (1989) J. Bacteriol.

171, 1870-1878.24. Nakasone, N., Yamashiro, T., Albert, M. J. & Iwanaga, M. (1993)

in Proceedings of the 26th Joint Conference, U.S.-Japan Co-operative Medical Science Program, Cholera and Related DiarrhealDiseases (National Institutes of Health, Bethesda, MD), pp.52-54.

25. Ward, H. M., Morelli, G., Kamke, M., Morona, R., Yeadon, J.,Hackett, J. A. & Manning, P. A. (1987) Gene 55, 197-204.

26. Gustafson, C. E., Chun, S. & Trust, T. J. (1994) J. Mol. Biol. 237,452-463.

27. Davis, M. A., Simons, R. W. & Kleckner, N. (1985) Cell 43,379-387.

28. Schulz, V. P. & Reznikoff, W. S. (1991) J. Mol. Biol. 221, 65-80.29. Pearson, G. D., Woods, A., Chiang, S. L. & Mekalanos, J. J.

(1993) Proc. Natl. Acad. Sci. USA 90, 3750-3754.30. Xiang, S.-H., Hobbs, M. & Reeves, P. R. (1994) J. Bacteriol. 176,

4357-4365.

Proc. Natl. Acad. Sci. USA 92 (1995)

Dow

nloa

ded

by g

uest

on

May

25,

202

0